SOME CONTEMl'OR.IRV .tPr.-ll^Cns LV PlIVSlCS-lll 271 



the (let'iKT levels of atomic electron-systems, as I described in the 

 secoml of these articles. The absorption in any particular substance 

 does not increase with an uninterrupted upward sweep; there are 

 occasional setbacks, each of which occurs at a critical frequency 

 where the radiation ceases to be able to extract electrons from a par- 

 ticular le\el. But though the lower-frequency ra\s cannot extract 

 the deeper electrt)ns of the atoms, they more than make up for it 

 by exjH'llinji the outer electrons in greater and greater abundance; and 

 when waNc-length 13A is reached, they can remove only the outermost 

 electrons or shift them from one orbit to another,'" but they perform 

 these actions so often that the beam is rapidly absorbed (even at O.oA, 

 0.01 mm. of lead is sufficient to abstract half its energy). 



Beyond 13A there is a region of well-nigh total eclipse. All we 

 know about it is derived from a few measurements by Holweck. 

 Accoriling to him, rays of wave-length 40.-\ lose half their energy in 

 tra\ersing half a millimetre of air at atmospheric density; at lOOA, 

 the siime proportion is consumed in a twentieth of a millimetre of air, 

 or in a quarter of a millimetre of hydrogen, the most tenuous of all sub- 

 stances; and even these are not the most absorbable rays. A sheet 

 of celluloid, .0001 mm. thick, which absorbs only 8^^ of the energy of 

 a beam of wave-length 40A and 'id^'c at lOOA, abstracts 9^% of the 

 energy at 250A. It actually absorbs 97.3% of a ray of wave-length 

 308A; but this may be the least penetrating radiation of the entire 

 scale, for the transmission apparently is a little greater at 400A 

 (although Holweck seems to distrust the reliability of the last result). 

 It must be admitted that the various beams of radiation on which 

 these measurements were made are not monochromatic, but comprise 

 each a continuous range of wave-lengths extending down to the quoted 

 value, which is the minimum. Since the beam is in every case filtered 

 through as many absorbing layers as possible before the final measure- 

 ment of transmission through the celluloid sheet is made, and those 

 remove preferentially the longer waves, it is probable that each 

 datum refers to a finite, yet comparatively narrow, band of wave- 

 lengths with its lower end at the specified value.* 



*Some of the absorbed encrg>' may be utilized in other ways, l)Ut there is no 

 known alternative mechanism. 



* The curve of Fig. 1, taken from Holweck's article, shows his data for the absorb- 

 ing power of celluloid plotted (logarithmically) against wave-length. .Ml the points 

 refer to wave-lenglhs l)ctween 40.-\ and 400.^ except the one marked "a," which 

 refers to the rays emitted by gaseous hydrogen liombardetl by electrons of energy 

 between 1,? and 38 volts (the transmission is the .same for every Ixjnibarding-voltage 

 within this range*. It is prolably a sort of "weighted-mean" value for the various 

 r.idiations of the Lyman series and possibly the secondary spectrum of hydrogen, 

 and the value 1 UO.A which Holweck assigns as its effective wave-length is probably 

 as good as any. The straight line on the left relates to nitrogen. 



